Development and initial validation of the institutional child abuse processes and coping inventory among a sample of Irish adult survivors of institutional abuse

Child Abuse & Neglect 33 (2009) 586–597

Contents lists available at ScienceDirect

Child Abuse & Neglect

Development and initial validation of the institutional child abuse processes and coping inventory among a sample of Irish adult survivors of institutional abuse夽 Roisin Flanagan-Howard a , Alan Carr a,∗ , Mark Shevlin b , Barbara Dooley a , Mark Fitzpatrick a , Edel Flanagan a , Kevin Tierney a , Megan White a , Margaret Daly a , Jonathan Egan c a b c

School of Psychology, University College Dublin, Belfield, Dublin 4, Ireland School of Psychology, University of Ulster, Magee campus, Londonderry, UK The Arches National Counselling Centre, HSE, Tullamore, Co. Offaly, Ireland

a r t i c l e

i n f o

Article history: Received 16 July 2007 Received in revised form 2 October 2008 Accepted 20 October 2008 Keywords: Institutional abuse Clerical abuse Abuse processes Coping with abuse

a b s t r a c t Objective: To develop a psychometric instrument to evaluate psychological processes associated with institutional abuse and coping strategies used to deal with such abuse. Methods: As part of a comprehensive assessment protocol, an inventory containing theoretically derived multi-item rational scales which assessed institutional abuse-related psychological processes and coping strategies were administered to 247 Irish adult survivors of institutional child abuse. Exploratory and confirmatory factor analyses were used to derive 6-factor scales, the reliability and validity of which were assessed. Results: Factor scales to assess the following constructs were developed (1) traumatization, (2) re-enactment, (3) spiritual disengagement, (4) positive coping, (5) coping by complying, and (6) avoidant coping. There were varying degrees of support for the validity of the scales with most support for the traumatization and re-enactment scales. Conclusions: The Institutional Child Abuse Processes and Coping Inventory (ICAPCI), particularly its traumatization and re-enactment scales, may be used in future research on adult survivors of institutional child abuse because they are currently the only scales that have been developed with this population to provide reliable and valid assessments of these constructs. Practice implications: The ICAPCI may be used, cautiously, to assess adult survivors of institutional child abuse. © 2009 Elsevier Ltd. All rights reserved.

Introduction Institutional abuse is distinguished from familial child abuse, in that it is perpetrated by adults working within the context of various types of institutions serving children in the community including residential care centers, schools, churches, and recreational facilities which may be managed by either secular or religious organizations (Gallagher, 2000). The present paper is specifically concerned with institutional abuse which occurred within the context of religiously affiliated residential institutions. Institutional abuse may involve physical, sexual or emotional maltreatment; it entails an abuse of power and a breach of trust; and typically, institutional abuse is an ongoing process rather than an isolated incident (Wolfe, Jaffe, Jette, &

夽 This study was funded by a grant from the Commission to Inquire into Child Abuse. ∗ Corresponding author. 0145-2134/$ – see front matter © 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.chiabu.2008.10.006

R. Flanagan-Howard et al. / Child Abuse & Neglect 33 (2009) 586–597

587

Poisson, 2003). In Ireland in recent years there have been many allegations of this form of institutional abuse. In response, the Irish Government set up the Commission to Inquire into Child Abuse (CICA). The research reported in this paper, is one element of a larger research project, which was commissioned by CICA. The overall objective of the research described in this paper was to develop a set of multi-item scales to assess psychological processes and coping strategies associated the experience of child abuse within the context of religiously affiliated institutions. There is some evidence that child abuse which occurs within religiously affiliated institutions has a significant impact on adult adjustment. For example, Wolfe, Francis, and Straatman (2006) found that in a sample of 76 adult male survivors of physical and sexual child abuse which occurred in residential religiously affiliated institutions, 88% had at some point in their lifetime met the diagnostic criteria for a DSM IV (APA, 2000) axis I disorder, and 59% presented with a current disorder. The most common conditions were posttraumatic stress, alcohol, and mood disorders. In addition on the Trauma Symptom Inventory (Briere, 1996), significant group mean elevations occurred on the trauma, dysphoria, depression, intrusive experiences, defensive avoidance, and dissociation scales. Over two thirds of the sample reported significant sexual problems in adulthood, and over half had a history of criminality. A variety of processes may account for the link between institutional child abuse on the one hand, adult adjustment on the other. Wolfe et al. (2003) propose that the long-term outcomes of institutional child abuse involves a set of distinctive psychological processes. These processes are traumatization, betrayal, stigmatization, disrespect for authority, and avoidance of reminders of abuse. These processes were identified in an extensive systematic literature review; a consideration of salient findings from clinical experience with over 450 cases of institutional abuse; and consultation with a panel consisting of abuse survivors, expert professionals and researchers from the child protection field. In this formulation, traumatization refers to the process of physical and psychological injury associated with abuse. Betrayal is the process that occurs when the child’s trusting relationship with the perpetrator is breached, and trust in others becomes diminished as a result. Stigmatization refers to the process of feeling shame, guilt, and humiliation as a result of being abused. Disrespect of authority is the process that occurs in response to being traumatized and betrayed by a perpetrator in an authoritative position. It may be accompanied by fear of authority figures and a sense of powerlessness. Avoidance of reminders of abuse is an avoidant coping strategy which abuse survivors use to inhibit recall of abusive experiences, which often occur as vivid distressing flashbacks when exposed to abuse-related cues. A further process which may account for the link between institutional child abuse on the one hand, adult adjustment on the other is religious disengagement. The research literature on child abuse perpetrated by religious leaders, such as priests and religious brothers, indicates that in addition to the processes identified by Wolfe et al. (2003), abuse survivors may also disengage from religious and spiritual beliefs and practices which they associate with their abusers (e.g., Bottoms, Shaver, Goodman, & Qin, 1995; Farrell & Taylor, 2000; Fater & Mullaney, 2000; McLaughlin, 1994; Wolfe et al., 2006). Religious disengagement is a ‘special case’, of the processes of avoiding reminders of abuse and disrespect of authority (identified in Wolfe et al., 2003 formulation outlined above), specific to cases of abuse perpetrated by religious leaders. Coping strategies, both adaptive and maladaptive, may also account for the link between institutional child abuse on the one hand, adult adjustment on the other. The research literature on stress and coping in children exposed to early childhood adversity suggests that children may use both functional and dysfunctional coping strategies to deal with institutional abuse (Luthar, 2003; Rutter, Quinton, & Hill, 1990). Functional coping strategies include social support, skill mastery, planning, and spiritual support. Dysfunctional coping strategies may include fully complying with the abusive regime or aggressively opposing it without due regard to the risks of further abuse entailed by this, or displacement of this anger and re-enactment of abuse on self or others. Excessive consumption of alcohol, drugs and food are other potentially dysfunctional coping strategies. The overarching objective of the present study was to develop a psychometric instrument to evaluate psychological processes associated with institutional abuse and coping strategies used to deal with it, informed by the literature briefly reviewed above. Within the context of this overall objective, there were a number of specific aims and hypotheses or research questions. The first aim was to establish a set of factorially valid factor scales which contained similar items for past and present versions of each scale. Past versions of scales would indicate respondents’ recollections of childhood experiences, while respondents current experiences, would be assessed by current versions of scales. An important benefit of establishing a set factor scales containing similar items for past and present versions of the instrument, would be that repeated measures analyses could be conducted to examine changes in participants’ perceptions of their status on factor scales from childhood to adulthood. We expected that a theoretically coherent factor structure would offer a parsimonious explanation for item response variability. The second aim was to evaluate the reliability of these scales, and the expectation was that they would show acceptable levels of internal consistency and inter-rater reliability. The third aim was to assess the discriminative validity of the scales. It was expected that they would have large and significant correlations with indices of institutional, but not family-based child abuse. The fourth aim was to determine the construct validity of the scales, and in this context it was expected that they would show large and significant correlations with indices of adult adjustment. The fifth aim was to further evaluate the construct validity of the scales by assessing the degree to which different types of institutional child abuse (sexual, physical, and emotional) were associated with different profiles on the scales. This was an exploratory analysis, not guided by specific hypotheses. The final aim was to further evaluate the construct validity of the scales by assessing profiles in subgroups of participants who differed in their current level of mental health. It was expected that those with poorer mental health would show more problematic profiles on the scales, and less improvement from past to present versions of the scales.

588

R. Flanagan-Howard et al. / Child Abuse & Neglect 33 (2009) 586–597

Methods Participants The study was commissioned by CICA (the Commission to Inquire into Child Abuse), a statutory body established by the Irish Government in 2000 to investigate and report on institutional abuse. Two hundred and forty-seven adult survivors of institutional abuse, recruited through CICA, participated in this study. All people who attended CICA before December 2005 and who reported institutional abuse were invited to participate in the study unless their whereabouts was unknown; they were resident outside Ireland and UK; they previously stated they did not want to participate in a research project; they previously stated they did not want to be contacted by CICA; they were known to be deceased; or they were known to be in poor health or to have a significant disability. The overall exclusion rate was 26%. The response rate for the study was 26%. Approximately 20% of CICA attenders participated in this study. One hundred and twenty-six were living and interviewed in Ireland. One hundred and twenty-one were living and interviewed in the UK. The sample included almost equal numbers of males (54.7%) and females (45.3%), with a mean age of 60 years (SD = 8.33; range = 40–83 years), and it had been 22–65 years since they had suffered institutional abuse. Thirty-four percent of participants were retired; 24% were unemployed; 27% were unskilled or semiskilled; and the remaining 15% had skilled or professional jobs. Forty-nine percent had never passed any state, college or university examination. Forty-nine percent were married. Participants in this study were probably better adjusted than other CICA attenders, and other survivors of institutional abuse, since older cases in poor health or with significant disabilities and who were homeless were excluded. Instruments Participants were interviewed with a standard assessment protocol which included the Institutional Child Abuse Processes and Coping Inventory (ICAPCI) and a set of standardized instruments to facilitate its validation. The ICAPCI was designed specifically for this study to evaluate psychological processes and coping strategies which the literature suggests are associated with the experience of institutional abuse and later life difficulties. The following processes and coping strategies were covered in a series of multi-item rational scales developed by the authors: (1) traumatization, (2) betrayal and loss of trust, (3) stigmatization, shame and guilt, (4) disrespect of authority, (5) re-enactment of abuse, (6) powerelessness, (7) coping by opposing, (8) coping by using alcohol drugs and food, (9) religious disengagement, (10) coping through spiritual support, (11) coping through planning, (12) coping through skill mastery, (13) coping through social support, (14) coping by complying, and (15) coping by avoiding reminders of abuse. Participants completed two versions the ICACPI. The first inquired about processes and coping strategies used in childhood while living in institutions, and the second inquired about the same processes and coping strategies in adulthood. For all items, 5 point response formats were used from 1 = never true, to 5 = very often true. The assessment protocol elicited information on demographic characteristics and history of institutional experiences as well as containing the following instruments: institutional and family versions of the Childhood Trauma Questionnaire (CTQ, Bernstein & Fink, 1998); the anxiety, mood and substance use modules of the Structured Clinical Interview for Axis I Disorders of DSM IV (SCID I, First, Spitzer, Gibbon, & Williams, 1996); the antisocial, borderline, avoidant and dependent personality disorder modules of the Structured Clinical Interview for DSM IV Personality Disorders (SCID II, First, Spitzer, Gibbon, & Williams, 1997); the Trauma symptom Inventory (TSI, Briere, 1996); the Global Assessment of Functioning Scale (GAF, Luborsky, 1962); and the UK version of the 100 item World Health Organization Quality of Life Scale 100 (WHOQOL, Skevington, 2005). Procedure The study was designed to comply with the code of ethics of the Psychological Society of Ireland. In addition, ethical approval for the study was obtained through the University College Dublin (UCD) Human Research Ethics Committee. Every effort was made to insure that the research interviews were carried out in a way that was minimally distressing for participants. Follow-up contact and support was offered to all candidates in collaboration with the National Counseling Service in Ireland and the Immigrant Counseling and Psychotherapy Service in the UK. Over the 6 months of data collection fewer than 5% of participants required referral for counseling. Between June and December 2005, CICA provided the research team with lists of participants, who were then recruited into the study by telephone. A team of 29 trained interviewers, all of whom had psychology degrees, conducted face-to-face interviews at multiple sites in Ireland and the UK. In most cases interviews were of about 2 h duration, although some took up to 5 h to complete. Participants were reimbursed for travel and subsistence expenses. Inter-rater reliability of all scales in the protocol was evaluated by conducting interviews with 52 participants in which pairs of interviewers from the interviewing team were present and each completed independent protocols for the same cases. Hardcopies of interview protocols were securely stored at UCD. Data were entered into an SPSS spreadsheet in an anonymized form, where they were verified and analyzed.

R. Flanagan-Howard et al. / Child Abuse & Neglect 33 (2009) 586–597

589

Results Development of ICAPCI factor scales and assessment of their factorial validity The first aim of the study was to establish a set of factorially valid factor scales which contained similar items for past and present versions of the ICAPCI. The related hypothesis was that a theoretically coherent factor structure would offer a parsimonious explanation for item response variability. The analytic strategy was to conduct initial exploratory principal components analyses (PCA) of past and present versions of the ICAPCI; use the results of these analyses to construct a set of factor scales; and then test the validity of these through confirmatory factor analysis. The sample size of 247 used satisfied the recommendation for exploratory factor analysis of a ratio of at least 5 participants per variable (Gorsuch, 1983) and exceeded the minimum sample size requirement of 200 for confirmatory factor analysis (Boomsma, 1982). Results of exploratory PCA of the ICAPCI, yielded similar, although not identical, factor solutions for past and present versions of the instrument. Items with high factor loadings from both of these analyses, and a small number of items of clinical and theoretical importance not included in the factor solutions of these analyses, were used to form the following 6-factor scales for evaluation through confirmatory factor analysis. Traumatization is a 14-item scale which assesses truamatization, betrayal and loss of trust, stigmatization, shame and guilt, and disrespect of authority. Re-enactment is a 9-item scale which assesses re-enactment of abuse; powerlessness, coping by opposing, and coping by using alcohol and drugs. Spiritual disengagement is a 5-item scale which assesses disengagement from religious practice and not using spiritual coping strategies. Positive coping is a 9-item scale which assesses coping through planning, skill mastery, and social support. Coping by complying is a 3-item scale which assesses coping by complying with the wishes of people in authority. Avoidant coping is a 3-item scale which assesses coping by avoiding thoughts and situations associated with abuse. Confirmatory factor analyses were conducted to evaluate the factorial validity of past and present versions of the 6 ICAPCI factor scales. Two confirmatory factor models were specified and estimated using LISREL 8.72 (Jöreskog & Sörbom, 2005a) and PRELIS 2.72 (Jöreskog & Sörbom, 2005b). Model 1 was for the past version and model 2 was for the present version of the ICAPCI. The goodness of fit for each model was assessed using the Sattora-Bentler scaled chi-square (S-B2 ), the Incremental Fit Index (IFI: Bollen, 1989), and the Comparative Fit Index (CFI: Bentler, 1990). Model 1 was a reasonable description of the sample data [S-B2 = 1292, df = 843, p < .001; RMSEA = .05 (90%CI .04–.05); CFI = .90; IFI = .90; SRMR = .08]. Table 1 contains the standardized factor loadings, all of which are significant (p < .05) with the exception of two items (I trusted everyone then; I thought I could do nothing to change my situation then). Model 2 was a reasonable description of the sample data [S-B2 = 1544, df = 842, p < .001; RMSEA = .06 (90%CI .05–.06); CFI = .90; IFI = .90; SRMR = .08]. All factor loadings, given in Table 1, are significant (p < .05). Thus, the confirmatory factor analyses supported the factorial validity of the 6-factor scales of the past and present versions of the ICAPCI. Reliability The second aim of the study was to evaluate the reliability of the 6-factor scales of past and present versions of the ICAPCI. To establish the reliability of the 6 ICAPCI factor scales, internal consistency alpha reliability coefficients (based on 247 cases) and inter-rater reliability coefficients (based on 52 cases) were calculated for past and present versions of each of the 6 ICACPI factor scales. Alpha reliabilities ranged from .51 to .87 (with 7 of the 12 alpha coefficients close to, or above .7) indicating moderate to good internal consistency reliability for all ICAPCI scales. For the past version of the ICAPCI, in rank order of magnitude, the alphas for the factor scales were traumatization: .75, coping by complying: .71, spiritual disengagement: .69, re-enactment: .62, positive coping: .62, and avoidant coping: .59. For the present version of the ICAPCI the alphas were traumatization: .87, spiritual disengagement: .78, re-enactment: .70, positive coping: .68, coping by complying: 56, and avoidant coping .51. Eleven of the 12 inter-rater reliability coefficients were above .7 indicating good inter-rater reliability for 11 scales and moderate inter-rater reliability for one scale. For the past version of the ICAPCI, in rank order of magnitude, the inter-rater reliabilities were positive coping: .99, traumatization: .97, re-enactment: .95, avoidant coping: .91, spiritual disengagement: .80, and coping by complying: .51. For the present version of the ICAPCI, the inter-rater reliabilities were coping by complying, .98, avoidant coping: .98, positive coping: .96, re-enactment: .94, traumatization: .90, and spiritual disengagement: .85. These results supported the reliability of the factor scales of the past and present versions of the ICAPCI. Discriminative validity: correlations of ICAPCI with indices of institutional and familial child abuse The third aim was to assess the discriminative validity of the ICAPCI factor scales. The related hypothesis was that ICAPCI scales would have large and significant correlations with indices of institutional, but not family-based child abuse. To assess the discriminative validity of the ICAPCI, Pearson product moment correlations were computed between all scales of the past and present versions of the ICAPCI on the one hand, and the total abuse scores of the institutional and family versions of the CTQ. Only correlations with an absolute value above .3 and significant at p < .01 were interpreted as indicating a moderate association between variables, that is, variables shared at least 9% of variance. For the 121 cases who completed the family

590

Table 1 Factor loadings for confirmatory factor analysis of the past and present forms of the Institutional Child Abuse Processes and Coping Inventory. Past version

T

R

S

P

C

A

Present version

T

R

PRESENT TRAUMATIZATION (Eigenvalue = 5.37, % variance = 38.38) Traumatization I feel hurt now .62 I feel frightened now .52 I feel sad now .62 I feel humiliated now .73

Betrayal and loss of trust I trusted everyone then (-) I felt betrayed then I cut myself off from other people then

Betrayal and loss of trust I trust everyone now (-) I feel betrayed now I cut myself off from other people now

Stigmatization shame and guilt I felt I was worthless then I felt I was dirty then I felt ashamed then I felt guilty and believed the abuse was my fault then Disrespect of authority I was angry at everyone in authority then I liked people in authority then (-) I respected everyone in authority then (-)

.04a .56 .43

Stigmatization shame and guilt I feel I am worthless now I feel I am dirty now I feel ashamed now I feel guilty and believe the abuse was my fault now Disrespect of authority I am angry with everyone in authority now I like people in authority now (-) I respect everyone in authority now (-)

.60 .56 .65 .37

.46 .19 .14

.04 .56 .43 .60 .56 .65 .37

.46 .19 .14

PAST RE-ENACTMENT (Eigenvalue = 2.56, % variance = 28.45) Re-enactment I felt the urge to attack or abuse other .55 people then I hurt other people then .31 I felt the urge to harm or injure myself then .46 I harmed or injured myself then .33

PRESENT RE-ENACTMENT (Eigenvalue = 2.86, % variance = 31.82) Re-enactment I feel the urge to attack or abuse other .55 people now I hurt other people now .31 I feel the urge to harm or injure myself now .46 I harm or injure myself now .33

Powerlessness I believed that my life was controlled by others then I thought I could do nothing to change my situation then Coping by opposing I planned revenge on my abusers then

.59

Powerlessness I believe that my life is controlled by others now I think I can do nothing to change my situation now Coping by opposing I am planning revenge on my abusers now

.57 .41

Coping by alcohol, drugs and food I drink alcohol to cope now I take other drugs to cope now

Coping by alcohol, drugs and food I drank alcohol to cope then I took other drugs to cope then

.19 .09a

PAST SPIRITUAL DISENGAGEMENT (Eigenvalue = 2.30, % variance = 45.99) Religious disengagement I had faith in God then (-) .83 I had faith in the church then (-) .77

S

.19 .09

.59 .57 .41

PRESENT SPIRITUAL DISENGAGEMENT (Eigenvalue = 2.36, % variance = 59.06) Religious disengagement I have faith in God now (-) .83 I have faith in the church now (-) .77

P

C

A

R. Flanagan-Howard et al. / Child Abuse & Neglect 33 (2009) 586–597

PAST TRAUMATIZATION (Eigenvalue = 4.22, % variance = 30.19) Traumatization I felt hurt then .62 I felt frightened then .52 I felt sad then .62 I felt humiliated then .73

I stopped praying then I only went mass then because I would be punished if I did not to Coping through spiritual support I prayed to God then, and that made the abuse bearable (-)

.35 .33

I do not pray now I do not go to mass now

.35 .33

.51

Coping through spiritual support I pray to God now, and that makes the abuse bearable (-)

.51

.38

.53

.32

.43 .51 .52

.16

I had a good friendship with an adult I could trust and this made the abuse bearable then

.30

I reminded myself that my mother or father was still alive, cared about me, and this made the abuse bearable then

.39

PAST COPING BY COMPLYING (Eigenvalue = 1.91, % variance = 63.61)

PRESENT POSITIVE COPING (Eigenvalue = 2.57, % variance = 28.59) Coping through planning Now I plan each day very carefully to avoid bad feelings and make good things happen (like having a laugh, getting well fed, and keeping warm) Now I still follow a plan to make sure my job suits me and makes my situation better When my partner and I are planning something important we take time to plan it very carefully Coping though skill mastery I put my energy into my work and that makes me feel better now I put my energy into sport or music and that makes me feel better now I put my energy into a skill that I can do well that makes me feel better now Coping through social support I have a good friendship with a close friend I can trust and this made the abuse bearable now (this friend is not my partner, husband or wife) I have a good friendship with a person I trust and look up to and this makes the abuse bearable now (this could be doctor or counselor but not a partner) I have a good relationship with my partner who I know cares about me and who I can tell my troubles to now and this makes the abuse bearable (a partner is a wife/husband/cohabite/lover) PRESENT COPING BY COMPLYING (Eigenvalue = 1.62, % variance = 54.00)

.38

.53

.32

.43 .51 .52

.16

.30

.39

R. Flanagan-Howard et al. / Child Abuse & Neglect 33 (2009) 586–597

PAST POSITIVE COPING (Eigenvalue = 2.31, % variance = 25.65) Coping through planning Then I planned each day very carefully to avoid abuse and make good things happen (like having a laugh, getting well fed, and keeping warm) When I was leaving school I followed a plan to get a job that would suit me and make my situation better When I was settling down with my partner, I waited for at least 6 months to make sure we were well suited to live together Coping though skill mastery I put my energy into my school work and that made me feel better then I put my energy into sports or music and that made me feel better then I put my energy into a skill that I could do well that made me feel better then Coping through social support I had a good friendship with a close friend I could trust and this made the abuse bearable then

591

592

Past version

T

R

Coping by complying I tried to behave well for the teachers/nuns/brothers/priests so I would not be punished then I was careful never to break a rule then I was careful always to show respect to the brothers, priests, nuns and teachers then (even if I didn’t feel respect) PAST AVOIDANT COPING (Eigenvalue = 1 .82, % variance = 60.54) Avoidance of reminders of abuse I avoided thinking about the abuse then I avoided situations that reminded me of abuse then I avoided people who reminded me of the abuse then

S

P

C

A

.68

.78 .57

.45 .73 .74

Present version

T

R

Coping by complying I try to behave well and fit in with people at work and in my family now to avoid conflict and arguments I am careful never to break a rule now I am careful always to show respect to people in authority now (even if I do not feel respect) PRESENT AVOIDANT COPING (Eigenvalue = 1.74, % variance = 56.78) Avoidance of reminders of abuse I avoid thinking about the abuse now I avoid situations that reminded me of abuse now I avoid people who remind me of the abuse now

S

P

C

A

.68

.78 .57

.45 .73 .74

Note: N = 247; T = traumatization; R = re-enactment; S = spiritual disengagement; P = positive coping; C = coping by complying; A = avoidant coping. Headings in bold lowercase are the names of ICAPCI rational scales containing the items beneath them. Headings in bold uppercase are the name of the 6-factor scales supported by confirmatory factor analyses. (-) means the item was reverse scored. 5 point response formats were used for all items from 1 = never true, 2 = rarely true, 3 = sometimes true, 4 = often true, 5 = very often true. a With the exception of two marked items all factor loading are statistically significant (p < .05).

R. Flanagan-Howard et al. / Child Abuse & Neglect 33 (2009) 586–597

Table 1 (Continued )

R. Flanagan-Howard et al. / Child Abuse & Neglect 33 (2009) 586–597

593

version of the CTQ, none of the correlations between the total abuse scale of the family version of the CTQ and the ICAPCI scales reached this magnitude and significance. In contrast, the total abuse scale of the institutional version of the CTQ had large (r > .3) and significant (p < .01) correlations with the traumatization (r = .47) and re-enactment (r = .39) scales of the past version of the ICAPCI, and with the traumatization scale (r = .32) of the present version of the ICAPCI. These results support the discriminative validity of the ICAPCI traumatization and re-enactment scales. Construct validity: correlations of the ICAPCI with indices of adult adjustment The fourth aim of the study was to determine the construct validity of the ICAPCI factor scales. The related hypothesis was that they would show large and significant correlations with indices of adult adjustment. To assess the construct validity of the ICAPCI, correlations were computed between all scales of the past and present versions of the ICAPCI on the one hand, and the following indices of adult adjustment on the other: the TSI total number of trauma symptoms, the total number of SCID I and II current and lifetime psychological disorders, the GAF global functioning rating, and the WHOQOL total quality of life score. Only correlations with an absolute value above .3 and significant at p < .01 were interpreted as indicating a moderate association between variables, that is, variables shared at least 9% of variance. Past and present versions of the traumatization and re-enactment scales had large (r > .3) and significant (p < .01) correlations with the total number of trauma symptoms on the TSI. The correlations were .32 and .40 for the traumatization and re-enactment scales of the past version of the ICAPCI, respectively; and .64 and .63 for the traumatization and re-enactment scales of the present version of the ICAPCI, respectively. The present version of the traumatization and re-enactment scales had large (r > .3) and significant (p < .01) positive correlations with the total number of SCID I and II disorders, and negative correlations with GAF global functioning ratings, and the WHOQOL total quality of life score. Traumatization correlated .32 with the total number of disorders, −.38 with GAF global functioning, and −.57 with quality of life. Re-enactment correlated .32 with the total number of disorders, −.44 with GAF global functioning, and −.57 with quality of life. The positive coping scale of the present version of the ICAPCI had a correlation of .36 with the WHOQOL total quality of life score. These results support the construct validity of the ICAPCI traumatization, re-enactment, and positive coping scales. ICAPCI profiles of groups who had suffered different types of child abuse The fifth aim of the study was to further evaluate the construct validity of the ICAPCI scales by assessing the degree to which different types of institutional child abuse (sexual, physical, and emotional) were associated with different profiles on the scales. This was an exploratory analysis without explicit hypotheses. To determine if cases who had suffered different types of institutional abuse had different ICAPCI scale profiles, the 247 cases were classified into three groups who reported that the worst form of institutional abuse to which they had been subjected was either sexual (N = 80), physical (N = 102), or emotional (N = 85). Participants’ statements about their worst experiences were classified as sexual abuse if the words sexual abuse or rape were mentioned, or if they reported genital, anal or oral sex, masturbation or other coercive sexual activities. Participants’ statements about their worst experiences were classified as physical abuse if physical violence, beating, slapping, or being physically injured were reported. Statements of actions involving humiliation, degradation, severe lack of care, withholding medical treatment, witnessing the traumatization of other pupils, and adverse experiences that were not clearly classifiable as severe sexual or physical abuse were classified as severe emotional abuse. Inter-rater agreement greater than 90% was achieved for a sample of 10% of statements. To aid profiling, all ICAPCI scales were scored so they each had a range of 1–5 by computing the mean item rating for each scale. A series of twelve 3 × 2, Groups × Time repeated measures ANOVAs, with post hoc tests where appropriate, were used to test for significant variation between groups on either past or present versions of each ICAPCI scales and to identify significant changes from past to present on each ICAPCI scale. To control for type 1 error associated with conducting multiple statistical tests, a significance level of p < .01 (rather than p < .05) was used in these ANOVAs. From Table 2 it may be seen that significant intergroup differences occurred for past re-enactment and present coping by complying. With past re-enactment, the mean of the sexual abuse group was significantly higher than those of the physical or emotional abuse groups. This indicates that those who reported that sexual abuse was the worst form of institutional abuse to which they had been subjected, also reported higher levels of re-enactment in childhood, compared with the other two groups. For present coping by complying, the mean of the physical abuse group was significantly higher than those of the sexual and emotional abuse groups. This indicates that those who reported that physical abuse was the worst form of institutional abuse to which they had been subjected, also reported a higher level of coping by complying in adulthood, compared with the other two groups. For all 6 ICAPCI scales significant time effects occurred. For traumatization and reenactment, mean scores decreased from the past to the present, but for spiritual disengagement, they increased. Positive coping mean scores increased from past to present, but coping by complying and avoidant coping mean scores decreased. There were no significant Group × Time interactions, indicating that there were no significant intergroup differences in the pattern of past and present scores. ICAPCI profiles of groups with different numbers of psychiatric diagnoses The final aim of the study was to further evaluate the construct validity of the ICAPCI scales by assessing profiles in subgroups of participants who differed in their current level of mental health. It was expected that those with poorer mental

594

Table 2 ICAPCI profiles for severe sexual, physical and emotional abuse. Scale

Traumatization

Spiritual disengagement

Positive coping

Coping by complying

Avoidant coping

Group 2

Group 3

Sexual abuse, N = 60

Physical abuse, N = 102

Emotional abuse, N = 85

Past

M SD

4.29 .71

4.20 .65

4.10 .59

Present

M SD

3.32 .88

3.31 .86

3.01 .93

Past

M SD

2.84 .77

2.44 .64

2.32 .63

Present

M SD

1.81 .74

1.68 .66

1.61 .60

Past

M SD

3.10 .74

2.95 .80

2.79 .77

Present

M SD

3.38 .78

3.20 .77

3.12 .85

Past

M SD

2.51 .97

2.41 .78

2.41 .76

Present

M SD

3.24 .98

3.14 .80

2.99 .93

Past

M SD

4.39 1.00

4.72 .48

4.54 .89

Present

M SD

3.45 1.04

3.94 .88

3.47 1.19

Past

M SD

4.08 1.22

3.94 1.25

3.74 1.20

Present

M SD

3.83 1.10

3.72 1.14

3.43 1.18

3 × 2 ANOVA Groups × Time

Time

Groups

.82

290.20***

2.62

2.90

241.61***

8.31***

.18

21.10***

3.62

.74

144.39***

.98

1.57

161.74***

7.66**

Group diffs.

1>2=3

2>1=3

.12

10.64**

2.65

Note: Group 1 reported predominantly sexual abuse as the worst thing they experienced in an institution. Group 2 reported predominantly physical abuse as the worst thing they experienced in an institution. Group 3 reported predominantly emotional abuse as the worst thing they experienced in an institution. Group diffs. = significant intergroup differences. To aid profiling all scales have a possible range of 1–5, which was obtained for each scale by summing items and dividing by the number of items. ** p < .01. *** p < .001.

R. Flanagan-Howard et al. / Child Abuse & Neglect 33 (2009) 586–597

Re-enactment

Group 1

R. Flanagan-Howard et al. / Child Abuse & Neglect 33 (2009) 586–597

595

health would show more problematic profiles on the scales, and less improvement from past to present versions of the scales. The 247 cases were classified into three groups. Group 1 contained 83 participants who had four or more current or lifetime diagnoses as assessed with the SCID I and SCID II. Group 2 contained 119 participants who had 1–3 current or lifetime diagnoses. Group 3 contained 45 participants who had no diagnoses. A similar approach to data analysis was used to compare ICAPCI profiles of the three groups as in the previous analysis. From Table 3 it may be seen that the three groups differed significantly in their mean scores on the past and present versions of the traumatization and re-enactment scales. On the past and present versions of the traumatization and re-enactment scales, group 1 obtained a significantly higher mean scores than groups 2 and 3. On the present versions of the traumatization and re-enactment scales, group 2 also obtained a significantly higher mean score than group 3. There was a significant Group × Time interaction for traumatization. Group 1 showed less reduction in traumatization from past to present (effect size = 1.27) than group 2 (effect size = 1.65) or group 3 (effect size = 1.67). For all 6 ICAPCI scales significant time effects occurred, replicating the effects shown in Table 2.

Discussion The first hypothesis – that a theoretically coherent factor structure would offer a parsimonious explanation for ICAPCI item response variability – was supported. Confirmatory factor analyses supported the 6-factor scale structure of past and present versions of the ICAPCI. The scales assessed (1) traumatization, (2) re-enactment, (3) spiritual disengagement, (4) positive coping, (5) coping by complying, and (6) avoidant coping. The second hypothesis that these scales would show moderate levels of internal consistency and test–re-test reliability was partially supported. The traumatization, re-enactment, and spiritual disengagement, scales had internal consistency and inter-rater reliability coefficients of at least .7 for their present versions and above or only marginally below.7 for their past versions. The positive coping, coping by complying, and avoidant coping scales showed moderate to good internal consistency (.51–.71) and inter-rater reliability (.51–.99) across past and present versions. The third hypothesis concerning the discriminative validity of the ICAPCI scales, was that they would have large and significant correlations with indices of institutional, but not family-based child abuse, was supported for the traumatization and re-enactment scales, but not the other ICAPCI scales. The fourth hypothesis concerning the construct validity of the ICAPCI scales, that they would show large and significant correlations with indices of adult adjustment, was supported for the traumatization, re-enactment and positive coping scales. The fifth aim of the study was to further evaluate the construct validity of the scales by assessing the degree to which different types of institutional child abuse were associated with different profiles on the scales. There was a unique association between severe sexual abuse and high scores on the ICAPCI past re-enactment scale, and also between severe physical abuse and high scores on the ICAPCI present coping by complying scale. The sixth and final hypothesis, concerning the construct validity of the ICAPCI scales, that those with poorer mental health would show more problematic profiles on the scales, and less improvement from past to present versions of the scales was partially supported. Those with poorest mental health obtained higher scores than other participants on past and present versions of the traumatization and re-enactment scales, and also reported less reduction in traumatization from past to present. The structure of the ICAPCI scales provides some validation for the clinical insights and theoretical positions that informed its development, and also suggested some interesting refinements of these positions. The ICAPCI assesses the psychological processes that Wolfe et al. (2003) proposed to characterize survivors of institutional abuse. The processes of traumatization, betrayal, stigmatization, and disrespect for authority all loaded on a single factor, assessed by the ICAPCI traumatization scale. The processes of re-enactment and powerlessness, combined with coping by opposing and coping by using alcohol, drugs, and food all loaded on a single factor assessed by the ICAPCI re-enactment scale. The process of religious disengagement identified in the literature on clerical abuse (Bottoms et al., 1995; Farrell & Taylor, 2000; Fater & Mullaney, 2000; McLaughlin, 1994; Wolfe et al., 2006) loaded on the ICAPCI Spiritual Disengagement scale. Coping by planning, skill mastery and social support, identified in the stress and coping literature as functional coping strategies (Luthar, 2003; Rutter et al., 1990) all loaded on the ICAPCI Positive Coping factor scale. The study had a number of limitations, three of which deserve particular mention. First, the factor structure of the ICAPCI was established on a single sample, and so requires replication in other samples to assess its stability. Second, there was no attempt to evaluate the accuracy with which people recalled the impact of abuse on them as young people in response to the past version of the ICAPCI, and so the results might well be accurate, or may be distorted by a range of factors including participants present psychological state and mental health. The fact that improvements occurred on all ICAPCI scales from past to present versions (noted in Tables 2 and 3) may reflect actual improvements in functioning, or participants’ bias to believe things are better now than in the past. A third limitation of the study was that no data were available on participants’ treatment seeking behavior. As a consequence it was not possible to compare profiles of treated and untreated cases, particularly their changes over time on past and present versions of the ICAPCI scales. Such analyses would have provided information of interest to clinicians who treat adults survivors or institutional child abuse.

596

Table 3 ICAPCI profiles for institutional abuse cases with four or more, three or less, or no psychiatric diagnoses. Scale

Traumatization

Spiritual disengagement

Positive coping

Coping by complying

Avoidant coping

Group 2

Group 3

3 × 2 ANOVA

4+ Diagnoses, N = 83

1–3 Diagnoses, N = 119

0 Diagnoses, N = 45

Groups × Time

Past

M SD

4.39 .52

4.16 .63

3.90 .78

9.19***

Present

M SD

3.73 .68

3.12 .83

2.60 .91

Past

M SD

2.87 .78

2.35 .57

2.21 .57

Present

M SD

2.16 .75

1.53 .49

1.23 .32

Past

M SD

3.01 .77

2.86 .78

2.95 .80

Present

M SD

3.29 .75

3.22 .78

3.06 .95

Past

M SD

2.31 .90

2.49 .76

2.52 .81

Present

M SD

2.88 .89

3.31 .85

3.01 .91

Past

M SD

4.64 .73

4.54 .80

4.56 .84

Present

M SD

3.50 1.01

3.73 1.08

3.78 1.06

Past

M SD

3.94 1.32

3.99 1.15

3.62 1.22

Present

M SD

3.82 1.10

3.70 1.06

3.17 1.35

Time 297.35***

Groups

Group diffs.

29.82***

1>2=3 1>2>3

1.58

214.63***

61.31***

1>2=3 1>2>3

1.12

14.16***

1.05

NS NS

3.10

113.41***

4.31

NS NS

2.49

140.28***

.31

NS NS

1.11

11.43**

3.97

NS NS

Note: Group 1 had four or more current or lifetime DSM IV TR diagnoses as assessed with the SCID I and 2. Group 2 had 1–3 current or lifetime diagnoses. Group 3 had no diagnoses. To aid profiling all scales have a possible range of 1–5, which was obtained for each scale by summing items and dividing by the number of items. ** p < .01. *** p < .001.

R. Flanagan-Howard et al. / Child Abuse & Neglect 33 (2009) 586–597

Re-enactment

Group 1

R. Flanagan-Howard et al. / Child Abuse & Neglect 33 (2009) 586–597

597

Practice implications The ICAPCI may be used in future research and (cautiously) in clinical practice to assess the following psychological processes and coping strategies in adult survivors of institutional abuse: (1) traumatization, (2) re-enactment, (3) spiritual disengagement, (4) positive coping, (5) coping by complying, and (6) avoidant coping. The items in Table 1 may be used to construct the questionnaire, and a 5 point response format may be used for all items from 1 = never true to 5 = very often true. Acknowledgements Thanks to the interviewing team; colleagues at the NCS in Ireland and ICAP in the UK for clinical support; colleagues at psychology departments at UCL, MMU, University of Aberdeen who provided interviewing facilities in the UK; colleagues at Right of Place, Aisling, and the London-Irish center for expert advice on survivor issues; Muriel Keegan for administrative support; Fred Lowe for liaison at CICA; and all participants for their generosity in taking part in the study. References American Psychiatric Association (APA). (2000). Diagnostic and statistical manual of the mental disorders (fourth edition-text revision, DSM-IV-TR). Washington, DC: American Psychiatric Association. Bentler, P. M. (1990). Comparative fit indices in structural models. Psychological Bulletin, 107, 238–246. Bernstein, D., & Fink, L. (1998). Childhood trauma questionnaire: A retrospective self-report. Manual. San Antonio, TX: The Psychological Cooperation. Bollen, K. A. (1989). Structural equations with latent variables. New York: Wiley. Boomsma, A. (1982). Robustness of LISREL against small sample sizes in factor analysis models. In K. G. Jöreskog, & H. Wold (Eds.), Systems under indirect observation: Causality, structure prediction (Part I) (pp. 149–173). Amsterdam: North Holland. Bottoms, B. L., Shaver, P. R., Goodman, G. S., & Qin, J. (1995). In the name of God: A profile of religion-related child abuse. Journal of Social Issues, 51, 85–111. Briere, J. (1996). Trauma symptom inventory. Odessa, FL: Psychological Assessment Resources. Farrell, D., & Taylor, M. (2000). Silenced by God—An examination of unique characteristics within sexual abuse by clergy. Counselling Psychology Review, 15, 22–31. Fater, K., & Mullaney, J. (2000). The lived experience of adult male survivors who allege childhood sexual abuse by clergy. Health Nursing, 21(3), 281–295. First, M., Spitzer, R., Gibbon, M., & Williams, J. (1996). Structured clinical interview for DSM-IV axis I disorders, clinician version (SCID-CV). Washington, DC: American Psychiatric Press. First, M., Spitzer, R., Gibbon, M., & Williams, J. (1997). Structured clinical interview for DSM-IV personality disorders (SCID-II). Washington, DC: American Psychiatric Press. Gallagher, B. (2000). The extent and nature of known cases of institutional child sexual abuse. British Journal of Social Work, 30, 795–817. Gorsuch, R. (1983). Factor analysis (2nd ed.). Hillsdale, NJ: Lawrence Earlbaum Associates. Jöreskog, K., & Sörbom, D. (2005a). LISREL 8.72. Chicago: Scientific Software Inc. Jöreskog, K., & Sörbom, D. (2005b). PRELIS 2.72. Chicago: Scientific Software Inc. Luborsky, L. (1962). Clinicians’ judgements of mental health. Archives of General Psychiatry, 7, 407–417. Luthar, S. (2003). Resilience and vulnerability: Adaptation in the context of childhood adversities. Cambridge: Cambridge University Press. McLaughlin, B. (1994). Devastated spirituality: The impact of clergy sexual abuse on the survivor’s relationship with God and the church. Sexual Addiction & Compulsivity, 1(2), 145–158. Rutter, M., Quinton, D., & Hill, J. (1990). Adult outcome of institution-reared children: Males and females compared. In L. Robins, & M. Rutter (Eds.), Straight and devious pathways from childhood to adulthood (pp. 135–157). New York: Cambridge University Press. Skevington, S. (2005). World Health Organization Quality of Life 100 UK Version. Bath, UK: WHO Centre for the Study of Quality of Life, University of Bath. Wolfe, D., Francis, K., & Straatman, A. (2006). Child abuse in religiously-affiliated institutions: Long-term impact on men’s mental health. Child Abuse & Neglect, 30, 205–212. Wolfe, D., Jaffe, P., Jette, J., & Poisson, S. (2003). The impact of child abuse in community institutions and organizations: Advancing professional and scientific understanding. Clinical Psychology: Science & Practice, 10(2), 179–191.